12 Nov
2007
12 Nov
'07
6:09 a.m.
On 11 Nov 2007 at 21:44, dwight elvey wrote:
It has been some time since I played with CRC's.
I once
had to write test to check CRC chips that used
the Chinese Remainder Theorem to make data correction.
I had to make sure that it corrected correctly by checking
each part of the error correction. That was many years
ago.
I've still got the docs for the WD chip that would do both ECC and
CRC. When I was securing data for computer forensics, I started with
two interlocked CRCs using different polynomials. Eventually, the
gummint expressed a preference for SHA hashes, so the CRC went the
way of the dodo. SHA was a LOT slower than CRC or even MD5.
Cheers,
Chuck